U.S. patent application number 10/571232 was filed with the patent office on 2007-07-19 for method for reaching a deployment decision.
Invention is credited to Mario Kroeninger, Robert Lahmann, Thomas Lich, Michael Schmid.
Application Number | 20070168098 10/571232 |
Document ID | / |
Family ID | 34384289 |
Filed Date | 2007-07-19 |
United States Patent
Application |
20070168098 |
Kind Code |
A1 |
Lich; Thomas ; et
al. |
July 19, 2007 |
Method for reaching a deployment decision
Abstract
A deployment decision for a vehicle restraint system is reached
as a function of vehicle dynamics data, at least one vehicle
transverse acceleration and one yaw rate about the longitudinal
axis of the vehicle being linked to one another as the vehicle
dynamics data to reach the deployment decision. The vehicle
transverse acceleration is then additionally subjected to a
threshold value decision for reaching the deployment decision, the
threshold value being set as a function of the integrated vehicle
transverse acceleration and the integrated yaw rate.
Inventors: |
Lich; Thomas; (Schwaikheim,
DE) ; Lahmann; Robert; (Nurnberg, DE) ;
Schmid; Michael; (Kornwestheim, DE) ; Kroeninger;
Mario; (Buehl, DE) |
Correspondence
Address: |
KENYON & KENYON LLP
ONE BROADWAY
NEW YORK
NY
10004
US
|
Family ID: |
34384289 |
Appl. No.: |
10/571232 |
Filed: |
September 16, 2004 |
PCT Filed: |
September 16, 2004 |
PCT NO: |
PCT/DE04/02068 |
371 Date: |
November 14, 2006 |
Current U.S.
Class: |
701/46 ;
701/45 |
Current CPC
Class: |
B60R 21/0132 20130101;
B60R 21/0133 20141201; B60R 2021/01327 20130101; B60R 2021/01322
20130101 |
Class at
Publication: |
701/046 ;
701/045 |
International
Class: |
B60R 22/00 20060101
B60R022/00 |
Foreign Application Data
Date |
Code |
Application Number |
Sep 25, 2003 |
DE |
10344613.3 |
Claims
1-6. (canceled)
7. A method for making a deployment decision for a restraint system
in a vehicle, comprising: detecting a vehicle transverse
acceleration and a yaw rate about the longitudinal axis of the
vehicle; linking the vehicle transverse acceleration and the yaw
rate about the longitudinal axis of the vehicle as vehicle dynamics
data; comparing the vehicle transverse acceleration to a first
threshold value, wherein the first threshold value is set as a
function of an integrated vehicle transverse acceleration and an
integrated yaw rate about the longitudinal axis of the vehicle; and
making the deployment decision based at least on the vehicle
dynamics data and the comparison of the vehicle transverse
acceleration to the first threshold value.
8. The method as recited in claim 7, wherein the first threshold
value is set as a function of a ratio of the integrated yaw rate
and the integrated vehicle transverse acceleration.
9. The method as recited in claim 7, wherein integrations for
deriving the integrated vehicle transverse acceleration and the
integrated yaw rate start as a function of the vehicle transverse
acceleration and end as a function of an integral value for the
vehicle transverse acceleration.
10. The method as recited in claim 8, wherein integrations for
deriving the integrated vehicle transverse acceleration and the
integrated yaw rate start as a function of the vehicle transverse
acceleration and end as a function of an integral value for the
vehicle transverse acceleration.
11. The method as recited in claim 7, further comprising: comparing
the yaw rate to a second threshold value, wherein the second
threshold value is set as a function of the integrated yaw rate and
the integrated vehicle transverse acceleration.
12. The method as recited in claim 8, further comprising: comparing
the yaw rate to a second threshold value, wherein the second
threshold value is set as a function of the integrated yaw rate and
the integrated vehicle transverse acceleration.
13. The method as recited in claim 11, wherein the second threshold
value is formed as a function of a quotient of the integrated yaw
rate and the integrated vehicle transverse acceleration.
14. The method as recited in claim 12, wherein the second threshold
value is formed as a function of a quotient of the integrated yaw
rate and the integrated vehicle transverse acceleration.
15. The method as recited in claim 8, wherein in the determination
of the ratio of the integrated yaw rate and the integrated vehicle
transverse acceleration, the integrated vehicle transverse
acceleration is subtracted from the vehicle transverse velocity at
the beginning of a lateral impact event.
16. The method as recited in claim 14, wherein in the determination
of the quotient of the integrated yaw rate and the integrated
vehicle transverse acceleration, the integrated vehicle transverse
acceleration is subtracted from the vehicle transverse velocity at
the beginning of a lateral impact event
Description
FIELD OF THE INVENTION
[0001] The present invention is directed to a method for reaching a
deployment decision for a restraint system in a vehicle.
BACKGROUND INFORMATION
[0002] A method for reaching a deployment decision for a restraint
system, e.g., for soil trips, is described in published German
patent document DE 101 49 112. Soil trips are understood to mean
situations in which the vehicle slides sideways after a skid and
then runs into a ground surface having a high coefficient of
friction, for example, an unpaved surface next to a roadway. If the
vehicle then slides to the right, for example, the tires on the
right side of the vehicle experience a severe deceleration which
then induces a torque on the vehicle on the unpaved surface. In
published German patent document DE 101 49 112, the deployment
decision is determined as a function of vehicle dynamics data,
i.e., a float angle in conjunction with a vehicle transverse
velocity and a vehicle tipping motion being used as the vehicle
dynamics data. The deployment decision is then reached through
appropriate threshold value comparisons.
[0003] Published international patent document WO 99/47384
describes reaching a deployment decision for a soil trip rollover
as a function of the yaw rate about the vehicle longitudinal axis,
a vehicle speed, and a vehicle transverse acceleration. The vehicle
transverse acceleration is compared to a fixed threshold value. It
is not possible to reach a deployment decision until this fixed
threshold value is exceeded.
SUMMARY
[0004] The method according to the present invention for reaching a
deployment decision for a restraint system has the advantage that
earlier deployment is made possible in the event of soil trips.
This is due to the fact that the vehicle transverse acceleration is
not only linked to the yaw rate, but is also compared to a
threshold value which is set as a function of the integrated yaw
rate and the integrated vehicle transverse acceleration. The
establishment of the threshold value results in a better adaptation
to accident conditions. The threshold may be set continuously or at
set time intervals. This threshold value decision may be made by
comparing a value pair, composed of the vehicle transverse
acceleration and the vehicle transverse velocity, to a
characteristic curve.
[0005] Analyses of soil trip vehicle tests have shown that the
vehicle transverse velocity has a crucial influence on the maximum
roll angle achieved, and thus on the rollover behavior of a
vehicle. In any case, for reaching a deployment decision, it is
advantageous for the signals to be detected from a yaw rate sensor
for rotations about the longitudinal axis of the vehicle, and to
link these signals with the signals from an acceleration sensor for
detecting acceleration in the transverse direction of the vehicle.
This results in a higher reliability and an earlier deployment
possibility for the restraint system. The vehicle transverse
acceleration is suitable, since, as described above, during a soil
trip a lateral deceleration occurs at the tires and initiates the
rollover.
[0006] It is also possible to use not only the vehicle transverse
velocity, but also the vehicle transverse acceleration, as well as
the yaw rate about the longitudinal axis of the vehicle, to enable
a high degree of reliability for the deployment decision and, at
the same time, a very early deployment decision.
[0007] It is particularly advantageous that the threshold value for
the vehicle transverse acceleration is generated as a function of
the quotient of the integrated yaw rate and the integrated vehicle
acceleration. This quotient is referred to as the rollover
susceptibility of the vehicle. The present invention makes use of
the following findings: When a body in motion due to its inertia is
decelerated by an externally acting force, the inert mass of the
vehicle experiences an inertial force. In the simplified assumption
of a rigid body, this inertial force may be represented by a force
vector acting on the center of gravity of the vehicle. This is
illustrated in FIG. 3. A vehicle 30 is subjected to inertial force
F.sub.inert, the vector of which points to the right. To the right
of the vehicle an obstruction 31 is also seen, there being a height
H1 between the center of gravity at which inertial force
F.sub.inert acts and obstruction 31.
[0008] The higher the center of rotation, thus for example the
upper edge of a roadway curb, which in this case is obstruction 31,
the lower the torque induced, under constant force, as a result of
the inertial force. In the second example in FIG. 3, vehicle 33
once again is subjected to an inertial force F.sub.inert which
points to the right, but now, height H2 is less since obstruction
32 is higher. Consequently, the torque is also lower. As a result,
the higher the center of rotation the greater the inertial force,
and thus also the deceleration measured in the vehicle--since the
vehicle mass may be assumed to be constant--must be in order to
cause the vehicle to overturn. When the height of the center of
rotation is equal to or greater than the height of the center of
gravity, the vehicle cannot be overturned at all. The acceleration
sensors in the vehicle measure an acceleration which allows
conclusions to be drawn concerning the magnitude and the direction
of the acting force, but not concerning the point of application of
the force. To obtain a measure of the effect of the laterally
acting force on the rollover behavior of the vehicle, the rollover
susceptibility of the vehicle S.sub.roll is computed as follows: s
roll = .DELTA..phi. x / .DELTA. .times. .times. v y = .intg. T 0
t_end .times. .omega. x .times. d t / .intg. T 0 t_end .times. a y
.times. d t ( 1 ) ##EQU1##
[0009] In this regard, the start and end points are generated by a
suitable calibration. One possible implementation is to define
starting time T.sub.0 as the time at which the acceleration has
exceeded a predefined threshold, and then to set end time T.sub.n
as the time when the integral over a.sub.y reaches a predefined
value.
[0010] It is possible that in the denominator of equation (1) the
integral over a.sub.y may be subtracted from the transverse
velocity present at the beginning of the soil trip on account of
sliding. An expression is then provided in the denominator for the
magnitude of speed which is instantaneously present in the
transverse direction, the reduction in speed resulting from the
impact being taken into account by the integral over a.sub.y. The
kinetic energy in the transverse direction of the vehicle may then
be easily computed. In this instance, a side impact is understood
to mean an impact on the side, for example on a curb, or also an
impact on the side as the result of the wheels digging into an
unpaved ground surface.
[0011] The computation of S.sub.roll may be refined in such a way
that the integral is formed only when an additional condition is
met, such as when the acceleration exceeds a minimum value. The
formation of S.sub.roll is then modified as follows: s roll =
.DELTA. .times. .times. .phi. x / .DELTA. .times. .times. v y =
.intg. T 0 t_end .times. ( f weight .times. .omega. x ) .times. d t
/ .intg. T 0 t_end .times. ( f weight .times. a y ) .times. d t ( 2
) ##EQU2## where the weighting function f.sub.weight=0 when the
additional condition, for example, the absolute value of
a.sub.y> a threshold, is not met, and f.sub.weight=1 in all
other cases. Thus, at any time during a rollover the rollover
susceptibility is determined, and the applicable threshold of a
base characteristic curve at this time modifies the applicable
formulas.
[0012] From the rollover susceptibility a variable g (S.sub.roll )
is then derived, which appropriately varies the threshold for
a.sub.y, which is generated according to a procedure. One
possibility for deriving variable g (S.sub.roll ) is provided by an
analytical formula or an additional characteristic curve (look-up
table) through which a variable g (S.sub.roll) is associated with
every value of S.sub.roll. The effect of g (S.sub.roll) may be, for
example, that an existing threshold of a.sub.y is increased by g
(S.sub.roll ), as follows: threshold (new)=threshold (old)+g
(S.sub.roll ), or is multiplied by g (S.sub.roll ), as follows:
threshold (new)=threshold (old)*g (S.sub.roll).
[0013] One example for the modification of a threshold by g
(S.sub.roll) is provided in FIG. 4.
[0014] Furthermore, it is advantageous for the yaw rate to be
compared to a threshold value which is likewise set as a function
of the integrated yaw rate and the integrated vehicle transverse
acceleration. Here as well, it is possible for the rollover
susceptibility to be used for establishing the threshold value for
the yaw rate, as described above. A base characteristic curve for
the threshold value is then also used for the yaw rate, the
threshold value being modified as a function of the rollover
susceptibility.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] FIG. 1 shows a block diagram of an example embodiment of the
system according to the present invention.
[0016] FIG. 2 shows a block diagram illustrating an example method
according to the present invention.
[0017] FIG. 3 illustrates the effects of various heights of
laterally positioned obstructions.
[0018] FIG. 4 shows the relationship between the rollover
susceptibility and the modified value for the thresholds.
DETAILED DESCRIPTION
[0019] Modern systems for sensing rollover events use
micromechanical yaw rate sensors which allow the rotational angle
to be computed by numerical integration. The combination of
information on the yaw rate and the rotational angle allows a
prediction of the rollover, and thus a deployment decision, which
is more robust and flexible than deploying via a fixed angle
threshold of an inclination sensor. Rollover sensing systems based
on yaw rate sensors thus allow deployment of irreversible
restraining means, such as pyrotechnic seat belt tensioners and
windowbags, in addition to the original applications of rollover
sensing, the deployment of a reversible roll bar in convertibles. A
classic rollover is induced when during straight-ahead driving the
vehicle is forced by conditions of the surroundings into a
z-directional motion, i.e., in the vertical direction, resulting in
a rotation of the vehicle. Typical examples of such situations
include sloping embankments next to the roadway, and ramps
typically provided with lateral guard rails. The lateral
accelerations which arise in such maneuvers are relatively low, and
the occupants are put into an "out of position" situation late, if
at all, so that the deployment of occupant protection systems is
not necessary until a relatively late point in time. In this
regard, "out of position" situation means that a passenger is not
in the seated position in which the restraining means provides the
optimum protection.
[0020] However, the matter is different for soil trip rollovers.
These are situations in which the vehicle slides sideways after a
skid and then runs into a ground surface having a high coefficient
of friction, such as an unpaved surface next to a roadway. If the
vehicle then slides to the right, the tires on the right side
experience a strong deceleration which then creates a torque on the
vehicle. The critical difference from the previously described
rollover is that, because of great lateral deceleration of the
vehicle, the passengers are put "out of position" very early. It is
therefore necessary to protect the passengers at a very early point
in time by activating the appropriate protective devices, such as
windowbags, for example, before injuries occur by impact with the
B-pillar or window. Such an early deployment is not possible with
current systems, at least not without incurring the risk of
unintentional deployment of the restraining means in many
non-deployment situations. The present invention provides a method
which allows an earlier deployment time in the case of soil
trips.
[0021] To allow such an early deployment time in the case of soil
trips, the present invention uses, in addition to the variables of
yaw rate and acceleration in the y and z directions, an
appropriately determined vehicle velocity in the y direction, i.e.,
the vehicle transverse velocity.
[0022] According to the present invention, the deployment decision
is reached in such a way that, in addition to linking the yaw rate
and the vehicle transverse acceleration, the vehicle transverse
acceleration is subjected to a threshold value decision, the
threshold value being set as a function of the integrated yaw rate
and the integrated vehicle transverse acceleration. The vehicle
transverse velocity may be used for this purpose.
[0023] The appropriately filtered acceleration in vehicle
transverse direction a.sub.y is particularly suited for the
threshold value decision, since a lateral acceleration on the tire
initiates the rollover. As confirmed by vehicle tests, in order to
roll over a vehicle, transverse acceleration a.sub.y must increase
with decreasing vehicle transverse velocity v.sub.y. Normally the
relationship is not linear, and is taken into account by the
threshold decision. Rather, as a function of the vehicle transverse
velocity, the critical transverse acceleration, i.e., the
transverse acceleration resulting in a rollover, shows a gradient
which becomes larger as the vehicle transverse velocity more
closely approaches the "critical sliding velocity" (CSV) from
higher speeds. The CSV is defined as the transverse velocity of the
vehicle below which a rollover of the vehicle due to physically
based principles, i.e., the energy balance, is impossible. The
exact shape of the characteristic curve depends on the type of
vehicle and the requirements for the system. However, in the
following example it may be assumed that the characteristic curve,
i.e., the absolute value of the critical transverse acceleration,
monotonically increases as a function of the vehicle transverse
velocity for decreasing values of v.sub.y.
[0024] Besides a.sub.y, the appropriately filtered yaw rates
.omega..sub.x about the longitudinal axis of the vehicle are also
suitable for the threshold value decision, which in this instance
is used as a link. The use of .omega..sub.x may be less intuitive,
since a lateral deceleration initiates the soil trip process.
However, analyses of corresponding vehicle tests have shown that
both .omega..sub.x and a.sub.y, with appropriate filtering, are
suitable as variables for a deployment decision.
[0025] By assuming a continuous, essentially exact determination of
the vehicle transverse velocity v.sub.y, after a starting condition
for the algorithm is met, the sensed values for a.sub.y,
.omega..sub.x, and v.sub.y are continuously, i.e., in each loop of
the algorithm, compared to the critical values stored in the form
of a characteristic curve. If the value pair (a.sub.y, v.sub.y) at
a time t is greater than the critical value of the characteristic
curve, the primary deployment condition is met. It must also be
ensured that the lateral acceleration also in fact induces a
rotation. This subject will be discussed further below. In addition
to a.sub.y, .omega..sub.x may also be compared to a threshold value
set as a function of v.sub.y, or the v.sub.y-dependent threshold
value may be modified as a function of .omega..sub.x.
[0026] In the following discussion it may be assumed that a.sub.y
is negative, i.e., is a deceleration, and that both v.sub.y and yaw
rate .omega..sub.x are positive. If it is assumed that a.sub.y is
determined by a sensor in the airbag control unit, the algebraic
sign depends on whether the soil trip occurs as the result of
lateral sliding to the left or right. Likewise, the algebraic sign
of v.sub.y depends on the convention used in the determination of
v.sub.y. The following method is provided for the implementation in
the microcontroller, i.e., in the processor in the airbag control
unit:
[0027] The absolute values for all variables v.sub.y, a.sub.y, and
.omega..sub.x are determined. In addition, an algebraic sign
ensures that v.sub.y, a.sub.y, and .omega..sub.xpoint in the same
direction as a condition for a rollover.
[0028] The lateral acceleration which results in tipping of the
vehicle is essentially determined by the position of the center of
gravity and the track width of a vehicle, and is determined by
computer using the static stability factor (SSF). Typical values
for automobiles and sport utility vehicles (SUVS) are in the
approximate range of SSF=1.0 to 1.7. The SSF corresponds to the
lateral acceleration, in units of g, necessary to tip over the
vehicle. A characteristic curve for |a.sub.y | at v.sub.y, will
therefore always have a value as the lowest deployment threshold
which is greater than the SSF value, in g, for the corresponding
vehicle. Depending on the ground surface, however, it is also
possible for a high acceleration to develop on all tires, i.e., not
only on the right or left tires, during lateral skids to the right
or left, respectively, so that, although the vehicle skids
sideways, no sufficiently large torque is induced to cause the
vehicle to tip. If one relies solely on a threshold value for
|a.sub.y | as a function of v.sub.y to be exceeded for the
deployment decision, in the worst case scenario this may result in
deployment under high vehicle transverse acceleration without
significant development of a tipping angle. To suppress deployment
in such cases, it is advantageous to link an additional deployment
condition to the yaw rate signal. The following methods provide a
possible implementation of the yaw rate signal being additionally
taken into account: [0029] a) As an additional deployment
condition, a threshold for the appropriately filtered yaw rate must
be exceeded. [0030] b) As an additional deployment condition, a
threshold for the integrated yaw rate, i.e., the resulting angle,
must be exceeded, it being advantageous to link the start of
integration to a threshold value for the yaw rate being exceeded.
[0031] c) Furthermore, the start of an integration of a yaw rate
may be linked to a threshold value for the vehicle transverse
acceleration being exceeded. In this case, the yaw rate is not
integrated unless the appropriately filtered vehicle transverse
acceleration is greater than a defined value. As an additional
deployment condition, it is then required that the resulting
integral, having the dimension of an angle, must exceed a threshold
value.
[0032] The above-described tasks do not result when a threshold
value for .omega..sub.x is considered as a function of v.sub.y.
However, even for driving maneuvers that are not relevant to soil
trips, such as for fast driving on narrow curves, very high yaw
rates may develop which then possibly may result in spurious
deployment. In this case, it is therefore advantageous to introduce
a threshold based on the sensor signal for the vehicle transverse
acceleration. Similarly to the above-described additional
deployment conditions based on the yaw rate signals, the following
implementation examples are described: [0033] a) As an additional
deployment condition the threshold for the appropriately filtered
vehicle transverse acceleration must be exceeded. [0034] b) As an
additional deployment condition a threshold for the integrated
vehicle transverse acceleration, i.e., the drop in velocity, must
be exceeded, it being advantageous to link the start of integration
to a threshold value for the vehicle transverse acceleration being
exceeded. [0035] c) Furthermore, an integration of the yaw rate may
be linked to a threshold value for the yaw rate being exceeded: in
this case, the vehicle transverse acceleration is not integrated
unless the appropriately filtered yaw rate is greater than a
defined value. As an additional deployment condition, it is then
required that the resulting integral, having the dimension of a
speed, must exceed a threshold value.
[0036] Thus, in any case it is advantageous for a deployment
decision to link the signals from a yaw rate sensor and an
acceleration sensor. Methods have been described heretofore in
which a primary deployment decision was made based on a
characteristic curve for a.sub.y and .omega..sub.x and then an
additional, less stringent deployment condition was based on a
plausibility check of the response of .omega..sub.x and a.sub.y. Of
course, an equivalent deployment decision for a.sub.y and
.omega..sub.x is also possible, i.e., characteristic curves may be
defined for both a.sub.y and .omega..sub.x, the deployment
decisions for which are appropriately linked, such as by a simple
logical AND. In addition, a.sub.y and .omega..sub.x may be suitably
processed (filtering and integration, for example) and linked.
[0037] FIG. 1 illustrates in a block diagram the system according
to the present invention. For the sake of simplicity, only the
components which are involved in the decision method according to
the present invention are shown, although it is possible for many
more components to be included in the system. A yaw rate sensor 10
for detecting yaw rate .omega..sub.x about the longitudinal axis of
the vehicle is connected to a first input of a processor 11. An
acceleration sensor 12 which detects accelerations in the
transverse direction of the vehicle is connected to a second input
of a processor 11. Restraining means 13, such as airbags, seat belt
tensioners, and roll bars are connected to an output of processor
11. Components 10, 11, and 12 may be located in a common control
unit. However, it is possible for sensors 10 and 12 to be situated
outside the control unit in which processor 11--which may be a
microcontroller, for example--is located, the sensors being
situated, for example, in a kinematic sensor platform. Sensors 10
and 12 may be connected to analog inputs of microcontroller 11. The
analog-digital conversion then occurs in microcontroller 11.
However, sensors 10 and 12 may each be digital sensors, which
already emit digital signals. Therefore, digital inputs are then
used for controller 11 to detect the sensor signals from yaw rate
sensor 10 and acceleration sensor 12.
[0038] Microcontroller 11 uses variables .omega..sub.x and a.sub.y
to make a deployment decision with respect to a rollover. Most
rollovers occur about the longitudinal axis of the vehicle. The
deployment decision is made as a function of threshold value
decisions concerning vehicle transverse acceleration a.sub.y and,
optionally, yaw rate .omega..sub.x. In this instance, this
threshold value is varied to account for various circumstances, the
various obstructions, which result in different heights of the
centers of rotation. The threshold value for a.sub.y is generated
from the quotient of integrated yaw rate .omega..sub.x and
integrated vehicle transverse acceleration a.sub.y, and a.sub.y is
then compared to this threshold value. If a.sub.y is greater than
the threshold value, the deployment decision is reached; if a.sub.y
is lower than the threshold value, the deployment decision is
suppressed.
[0039] FIG. 2 illustrates in a block diagram the sequence of the
method according to the present invention. Vehicle transverse
acceleration a.sub.y is detected by acceleration sensor 12 in block
20. Vehicle transverse acceleration a.sub.y is integrated in block
21 and compared to a threshold value in block 24, which threshold
value is determined as a function of the integrated vehicle
transverse acceleration and the integrated yaw rate from block 23.
The result of the threshold value comparison is recorded in block
26 in order to then reach the deployment decision. Yaw rate
.omega..sub.x is detected by sensor 10 in block 22. The yaw rate is
likewise integrated in block 23 and subjected to a threshold value
comparison in block 25, this threshold value also being generated
as a function of the integrated yaw rate and the integrated vehicle
transverse acceleration. As illustrated, the threshold value
generation is performed by forming a quotient, so that the rollover
susceptibility sets the threshold value in each case. The result of
the threshold value comparison of yaw rate .omega..sub.x with its
threshold value in block 25 is recorded in block 27, and is
available for further processing.
[0040] FIG. 3 shows two typical situations involving vehicle
torques. On the left, vehicle 30 is subjected to an inertial force
F.sub.inert to the right. The inertial force acts on the center of
gravity of the vehicle. Therefore, the arrow representing the
inertial force is drawn in at that point. Vehicle 30 therefore
moves to the right against obstruction 31. A torque is thus induced
by obstruction 31 which results from the product F.sub.inert*H. H
is the vertical distance between the upper edge of the obstruction
(center of rotation) and the center of gravity of the vehicle.
Comparing the left-hand and right-hand illustrations in FIG. 3,
Hi>H2, so that in the right-hand illustration a greater force,
i.e., deceleration, must act on the vehicle to generate the same
torque when the vehicle begins to tip.
[0041] In FIG. 4, rollover susceptibility S.sub.roll is plotted in
arbitrary units on the x-axis. Modification factor g (S.sub.roll)
is likewise plotted in arbitrary units, on the y-axis. A curve 40
describes the relationship therebetween. In this instance the
relationship is empirical. Curve 40 may alternatively be linear or
exponential.
* * * * *